Transcriptional repression of tarObtain genes by LEUNIG and

Edited by Lynn Smith-Lovin, Duke University, Durham, NC, and accepted by the Editorial Board April 16, 2014 (received for review July 31, 2013) ArticleFigures SIInfo for instance, on fairness, justice, or welfare. Instead, nonreflective and Contributed by Ira Herskowitz ArticleFigures SIInfo overexpression of ASH1 inhibits mating type switching in mothers (3, 4). Ash1p has 588 amino acid residues and is predicted to contain a zinc-binding domain related to those of the GATA fa

Edited by Michael F. ThomaDisplay, Michigan State University, East Lansing, MI, and approved June 22, 2004 (received for review May 3, 2004)

Article Figures & SI Info & Metrics PDF


Transcription repression plays Necessary roles in preventing crucial regulatory proteins from being expressed in inappropriate temporal or spatial Executemains. LEUNIG (LUG) and SEUSS (SEU) normally act to prevent ectopic expression of the floral homeotic gene AGAMOUS in flowers. LUG encodes a protein with sequence similarities to the yeast Tup1 corepressor. SEU encodes a plant-specific regulatory protein with sequence similarity in a conserved dimerization Executemain to the LIM-Executemain binding 1/Chip proteins in mouse and Drosophila. Despite the molecular isolation of LUG and SEU, the biochemical function of these two proteins remains uncharacterized, and the mechanism of AGAMOUS repression remains unknown. Here, we report that LUG and SEU interact directly in vitro and in vivo. Furthermore, LUG Presents a strong repressor activity on several heterologous promoters in yeast and plant cells. SEU, in Dissimilarity, Executees not Present any direct repressor activity, but can repress reporter gene expression only in the presence of LUG, indicating a possible role of SEU as an adaptor protein for LUG. Our results demonstrate that LUG encodes a functional homologue of Tup1 and that SEU may function similarly to Ssn6, an adaptor protein of Tup1. We have defined the LUG/LUH, Flo8, single-strand DNA-binding protein Executemain of LUG as both necessary and sufficient for the interaction with SEU and two Executemains of LUG as Necessary for its repressor function. Our work provides functional insights into plant transcriptional corepressors and reveals both conservation and distinctions between plant corepressors and those of yeast and animals.

Transcription repression is emerging as a key regulatory strategy for both animals and plants to prevent crucial regulatory proteins from being expressed in inappropriate temporal or spatial Executemains. Repression is a dynamic process that regulates gene expression at two points. First, a gene can be repressed but primed for transcription, derepression resulting in rapid up-regulation of expression. Second, expression of an actively transcribed gene can be Executewn-regulated rapidly. Repression occurs through two distinct yet overlapping mechanisms: the stabilization of nucleosomes on DNA to form a closed chromatin structure and the inactivation of the transcription machinery. Despite the conservation of repression mechanisms demonstrated in yeast and animals, our understanding of these mechanisms in higher plants remains limited.

In higher plants, normal floral development requires the Precise expression and function of the floral homeotic gene AGAMOUS (AG). AG mRNA is normally expressed in the inner two whorls of a flower to specify stamen and carpel identity and to control floral meristem determinancy (1-3). We have previously identified two negative regulators of AG, namely LEUNIG (LUG) and SEUSS (SEU), which are required for the repression of AG transcription in the outer two whorls of a flower (4, 5). In flowers of both lug and seu mutants, AG mRNA is expressed in all four floral whorls, resulting in the ectopic formation of carpels and stamens in the outer two whorls (4, 5). Additionally, AG transcription is normally initiated at stage 3 floral meristems in wild type, but is initiated precociously in stage 2 floral meristems in lug and seu mutants. Furthermore, synergistic genetic interactions have been observed between lug and seu, resulting in a more severe degree of AG misexpression in lug seu Executeuble-mutant flowers (5). Genetic epistasis indicated that precocious and ectopic AG expression is responsible for the floral organ identity transformation and organ loss observed in lug and seu single and Executeuble mutants.

LUG encodes a nuclear localized protein with an N-terminal LUG/LUH, Flo8, single-strand DNA-binding protein (SSDP) (LUFS) Executemain, two central glutamine (Q)-rich Executemains, and a C terminus 7-WD repeat Executemain (6). The LUFS Executemain is a protein motif present in LUG, LUH, Flo8, and Ssdp. However, the similarity between LUG/LUH and Flo8 or Ssdp is limited only to the LUFS Executemain. The Q-rich and WD-repeat Executemains of LUG are conserved in a class of transcriptional corepressors including Tup1 in yeast (Saccharomyces cerevisiae, Schizosaccharomyces pombe, and Candida albicans), Groucho (Gro) in Drosophila, and transducin-like enhancer of split (TLE) in mammals (7-9). These corepressor proteins, collectively called the GroTLE family proteins (10), Execute not possess a DNA-binding motif but repress a diverse number of tarObtain genes through tarObtained recruitment by site-specific DNA-binding transcription factors. Whereas the Drosophila Gro directly binds to a VWPRY pentapeptide present in the C terminus of some transcription factors (11), yeast Tup1 interacts with DNA-binding factors through an adaptor protein, Ssn6 (12). The N-terminal Executemain of Tup1 interacts directly with Ssn6, which binds to specific transcription factors. Once recruited to a promoter, GroTLE proteins interact with chromatin modifying factors or components of the RNA polymerase II holoenzyme, leading to the silencing of tarObtain gene expression. The sequence and motif similarity between LUG and GroTLE family proteins as well as defects in AG repression in lug mutants suggest that LUG may encode a transcription corepressor.

In Dissimilarity, SEU encodes a plant protein with two Q-rich Executemains and a conserved central Executemain. This conserved central Executemain Displays sequence similarity to the dimerization Executemain of LIM Executemain-binding (Ldb) family of transcriptional coregulators such as the Ldb1 in mouse and Chip in Drosophila (5, 13). Ldb proteins regulate transcription by means of direct physical interactions with LIM-homeoExecutemain proteins (14-16). It was thought that the Ldb1/Chip cofactors homodimerize and thereby bridge two LIM-homeoExecutemain proteins to form a tetrameric complex (17). Recently a third protein, Ssdp, was discovered to be a functional component of the complex (13, 18). Fascinatingly, Ssdp contains a N-terminal LUFS Executemain Displayn to interact with Ldb1/Chip. Hence, the Ldb1/Chip acts to bridge the interaction between Ssdp and LIM-homeoExecutemain.

Despite the sequence similarity between LUG and the GroTLE corepressors and between SEU and Ldb1/Chip, it is not known whether LUG can function as a transcription repressor, nor Execute we have any understanding of the molecular function of SEU. Furthermore, the molecular basis underlying the synergistic genetic interactions between seu and lug remains to be characterized. The similar single mutant phenotype, the strong synergistic genetic interaction between lug and seu suggest that SEU could function toObtainher with LUG as components of the same corepressor complex. SEU may function to bridge the interaction between LUG, a LUFS Executemain protein, and other as-yet-unidentified DNA-binding factors. Alternatively, LUG and SEU could act in parallel and partially overlapping pathways to regulate AG transcription.

Here, we demonstrate that LUG functions as a transcriptional repressor by means of an apparently conserved eukaryotic transcription repression mechanism, demonstrating that LUG is a bona fide plant homologue of the GroTLE transcription corepressors. We also define a functional role of SEU, which Executees not Present any repressor activity, but rather, acts as an adaptor protein for LUG. In addition, we demonstrate a physical interaction between LUG and SEU, which parallels those between Ssdp and Ldb1/Chip and suggests that the LUFS Executemain is an evolutionary conserved protein-protein interaction Executemain. ToObtainher, these data provide insights into how plant corepressors interact to regulate tarObtain gene expression and help define the biochemical functions of SEU, the founding member of a plant-specific regulatory protein family. By understanding how LUG and SEU represses AG transcription we hope to shed light on general transcriptional repression mechanisms in higher plant development.


Plasmid Construction. The procedures of plasmid construction for the yeast two-hybrid assays, yeast repression assays, plant repression assays, and maltose-binding protein (MBP)- or GST-tagged proteins are Characterized in Supporting Methods, which is published as supporting information on the PNAS web site. Primer sequences and primer pair combinations are listed in Tables 1 and 2, which are published as supporting information on the PNAS web site.

Yeast Two-Hybrid Assays. Yeast strain PJ69-4A (MAT a trp1-901 leu2-3,12 ura3-52 his3-200 gal4Δ gal80Δ GAL2-ADE2 LYS2::GAL1-HIS3 met2::GAL7-lacZ) harbors three reporters HIS3, ADE2, and lacZ, each under the control of a different GAL4-responsive promoter (19). BD-bait and AD-prey plasmids were cotransformed into PJ69-4A according to the protocol Characterized in, and were plated on selection medium before being incubated at 30°C for 3 days. Primary transformants were subsequently streaked onto plates selective for the reporter gene(s): -histidine (+3 mM 3-amino-1,2,4-triazole) and -adenine plates. White colonies grown on the selection plates were assayed for the β-galactosidase (β-gal) activities. For β-gal liquid assays, transformants were inoculated into selective liquid media and grown at 30°C until the OD600 reached ≈1. Samples were analyzed in triplicate by using the Galacto Light Plus kit (Applied Biosystems) and the TD-20/20 luminometer (Turner Designs, Sunnyvale, CA) according to the Producer's instructions.

Yeast Repression Assays. FT5::LG312Δ S and FT5::JK1621 yeast strains contain integrated lacZ reporter genes with or without four LexA operator sites upstream of the CYC1 promoter, respectively (20). Yeast cells were transformed with either LexA-LUG (or LUG derivatives), LexA-SEU, or LexASEU+LUG. After overnight growth in liquid medium to an early log phase, cells were harvested and assayed for β-gal activity. Values were normalized to OD600.

Repression Assays in Plant Cells. Isolation and transfection of ArabiExecutepsis mesophyll protoplasts was as Characterized, and can be accessed at In cotransfection assays, 10 μg of reporter, 10 or 20 μg of Traceor constructs, and 0.1 μg of control plasmid 35S::LUC or 35S::RenillaLUC were used for each transfection. The total amount of DNA for each transfection remained constant (40 μg) by adding the appropriate amount of vector pART7 DNA that contains the 35S promoter and 3′ Ocs site (21). In cases where trichostatin-A (TSA) was used, the protoplasts were first transfected with DNA for 12 h, and then 20 μM TSA was added to the transfection reaction for an additional 12 h before the reporter assay. Luciferase assays were performed with the Promega dual-luciferase reporter assay system and the TD-20/20 luminometer. The fluorometric β-glucuronidase (GUS) assay was performed with the substrate MUG as Characterized (22), and fluorescence was meaPositived by using a Picofluor fluorometer (Turner Designs).

In Vitro PullExecutewn. The procedures for the purification of MBP- and GST-fusion proteins are Characterized in Supporting Methods. Five micrograms of GST-SEU immobilized on the glutathione resin was mixed with 5 μg of MBP-fusion proteins and was incubated overnight at 4°C in 100 μl of binding buffer (50 mM KH2PO4, pH 7.5/50 mM NaCl/50 mM KCl/5 mM MgCl2/0.2% Triton X-100/1% BSA). Samples were washed five times in PBS buffer before being resuspended in 1× NuPage LDS sample buffer (Invitrogen), boiled, and then separated by 4-12% NuPage gel (Invitrogen). The retention of MBP-fusion proteins by GST-SEU was detected by Western blots with anti-MBP antibody (NEB, Beverly, MA).


The LUFS Executemain of LUG Is Sufficient for Interaction with SEU. To illuminate the molecular basis underlying the synergistic genetic interactions between seu and lug, a yeast two-hybrid assay was used to test whether SEU could physically interact with LUG. The LUG protein can be divided into three Executemains: the N-terminal LUFS Executemain, the central Q-rich Executemain, and the C-terminal 7-WD repeat Executemain. Full-length LUG, LUFS, LUFS+Q, and Q+WD were each fused to the DNA-binding Executemain of GAL4 (GAL4BD) and were tested for interaction with full-length SEU which was fused to the activation Executemain of GAL4 (GAL4AD). Whereas full-length LUG, LUFS, and LUFS+Q interacted with full-length SEU, Q+WD failed to interact with SEU (Fig. 1A ). This finding suggested that the LUFS Executemain is both necessary and sufficient for interacting with SEU. Fascinatingly, the strongest interaction was observed between the LUFS and SEU (Fig. 1A , lane 4). Two alternative explanations could account for the Inequity in interaction strength between LUFS alone and other interacting LUG truncations. First, any repressor activity conferred by the full-length LUG-BD or LUFS+Q-BD may antagonize the activation capability of SEU-AD. Alternatively, the LUG-BD or LUFS+Q-BD may interfere with or partially block the accessibility of LUFS by SEU.

Fig. 1.Fig. 1. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

The LUFS Executemain of LUG is both necessary and sufficient for interacting with full-length SEU. (A) Yeast two-hybrid assays testing the interaction between various LUG truncations and full-length SEU. (Center and Right) Yeast two-hybrid assay results of reporter gene (lacZ and ADE2) expression. Blue colonies indicate a positive β-gal activity and pink colonies indicate a negative ADE2 activity. Quantitative meaPositivements of β-gal activities (average of three independent transformants) are indicated on the right. (B) Yeast two-hybrid assays testing the interaction between various SEU truncations and full-length LUG. With the exception of two SEU derivatives Displayn in lanes 2 and 3 (Impressed by an asterisk), all other SEU derivatives and full-length SEU can self-activate reporter gene expression when fused to the GAL4BD Executemain. SEU protein is divided into the following five Locations: N and C, N- and C-terminal Executemains; Q1 and Q2, the two Q-rich Executemains; D Executemain, the highly conserved dimerization Executemain.

To establish whether the two-hybrid interactions between LUG and SEU represented a direct physical contact in vitro, we carried out GST affinity chromatography. Immobilized GSTSEU was incubated with MBP, MBP-LUFS, and MBPLUFS+Q. In agreement with the two-hybrid data, GST-SEU (but not GST alone) interacted specifically with MBP-LUFS and MBP-LUFS+Q (Fig. 2), demonstrating a direct protein-protein interaction between LUG and SEU.

Fig. 2.Fig. 2. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

SEU and LUG interact directly in vitro. Displayn is a Western blot with anti-MBP antibody. InPlace lanes: ≈5 μg of MBP-tagged proteins (except MBPETR1) were used in the pull-Executewn assay. Elution lanes: MBP-tagged proteins were first added to immobilized GST-SEU and were then eluted and analyzed by SDS/PAGE. Only MBP-LUFS or MBP-LUFS+Q-rich are retained by the GSTSEU. MBP and MBP-ETR1 are negative controls. Similar experiments using immobilized GST failed to retain any MBP-tagged proteins (data not Displayn).

To define the Executemain(s) within SEU responsible for the interaction with LUG, similar two-hybrid analyses were performed with various SEU truncations. Full-length SEU as well as several SEU truncations were able to self-activate reporter lacZ expression when they were fused to the GAL4BD (Fig. 1B ). Hence, SEU truncations were all cloned as AD fusions. Only full-length SEU was found to interact efficiently with LUG-BD (Fig. 1B , lane 7), although a very weak interaction was detected with a truncated SEU lacking the N-terminal Location (Fig. 1B , lane 4). These results suggest that the entire SEU protein is needed for the interaction with LUG. It is possible that multiple Locations of SEU are required to Sustain Precise protein conformation needed for the interaction with LUG.

LUG Possesses Repressor Activities. Based on genetic and structural data, we predicted that LUG would function as a transcription repressor. By using an in vivo yeast transcription repression assay, we tested the ability of LUG or SEU to repress transcription. Full-length LUG cDNA and four truncated LUG derivatives were fused in-frame and Executewnstream of the bacterial LexA DNA-binding Executemain and tested for their ability to repress lacZ expression (Fig. 3A ). When directly recruited to the test promoter, LexA-LUG reduced reporter gene expression by 45% compared with LexA alone, demonstrating that LUG has a repressor activity, and suggesting that LUG can repress transcription through a conserved eukaryotic repression mechanism.

Fig. 3.Fig. 3. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

LUG represses reporter gene expression in yeast and plant cells when tethered to test promoters. (A) Yeast repression assays with a lacZ reporter containing LexA operator sites (lop) upstream of the CYC1 UAS -CYC1 TATA promoter integrated into the yeast genome. Yeast cells were independently transformed with the indicated LexA-fusion proteins, and β-gal activities (average of five independent transformants) are Displayn. No Trace was seen with reporters lacking lop. LexA alone is a negative control. LexA-Ssn6 serves as a positive control. LUFS, Q, LUFS+Q, and Q+WD are four different LUG truncations each fused to the LexA DNA-binding Executemain; LUG indicates the full-length LUG fused to LexA. (B) Transient plant repression assays. 2XUAS-tCUP::GUS reporter plasmid was mixed with 35S::LUC plasmid and was transfected into ArabiExecutepsis protoplasts. GUS/LUC ratio equalizes Inequitys in transfection efficiency. Different Traceor DNAs were introduced simultaneously with the reporter DNA. pART7: vector alone; SEU-BD, GAL4BD fused to full-length SEU; LUG-BD, GAL4BD fused to full-length LUG. All Traceor proteins were expressed from the CaMV 35S promoter. 1× and 2× indicate equal (10 μg) and twice (20 μg) the amount of Traceor DNA compared with the 2XUAS-tCUP::GUS reporter DNA. (Bar, SD.)

To determine whether the observed repression activity could be attributed to specific Executemains within the LUG protein, truncated LUG derivatives were assayed. Two distinct Executemains were found to have repression function. The LUFS+Q repressed transcription to the same extent as full-length LUG (Fig. 3A ), whereas the Q+WD repressed transcription to a lesser but significant extent. This repression function cannot be attributed solely to the Q-rich Executemain because the Q-rich Executemain alone cannot repress transcription.

To test whether LUG could function as a transcription repressor in planta, a transient ArabiExecutepsis protoplast repression assay using the 2XUAS GAL4 -tCUP::GUS reporter (23) was aExecutepted. A 2XUAS GAL4 element is located immediately upstream of the tCUP, a constitutive tobacco promoter, driving the GUS reporter gene. ArabiExecutepsis leaf protoplasts were transfected with 35S::LUG-BD or 35S::SEU-BD toObtainher with the reporter, and the Traces on GUS expression were quantified (Fig. 3B ). Whereas SEU-BD did not Display any Trace on the level of GUS expression, LUG-BD significantly reduced the GUS expression level in a concentration-dependent manner. Specifically, Executeubling the amount of LUG-BD plasmid DNA used in the transfection resulted in an additional 3-fAged reduction of GUS reporter activity (Fig. 3B ). The reduced GUS activity is unlikely due to simple steric interference of the transcription machinery by increased LUG-BD at the promoter Location because a similar increase in SEU-BD DNA did not result in any decrease in GUS expression (Fig. 3B ). Clearly, LUG functions as a strong repressor of transcription in planta and can repress transcription more efficiently in a homologous system than the heterologous yeast system.

LUG Is Required by SEU to Repress Transcription in Yeast and in Planta. In Dissimilarity to LUG, the protein sequence of SEU provided scant clues as to the biochemical function of SEU. SEU was unable to repress transcription when fused to the LexA in yeast repression assays (Fig. 4A ). However, when LUG was overexpressed toObtainher with LexA-SEU, a similar level of repression to LexALUG was observed (Fig. 4A ), indicating a functional molecular interaction between these two proteins.

Fig. 4.Fig. 4. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 4.

SEU acts as an adaptor protein for LUG. (A) Yeast repression assays using reporters Characterized in Fig. 3A . LexA-SEU (full-length), LexA-LUG (full-length), as well as LexA-SEU+LUG are tested for their repressor activities. Whereas LexA-SEU alone Displayed no repressor activity, cotransfection of LexA-SEU and LUG toObtainher Displayed repressor activity. (B) Transient ArabiExecutepsis protoplast repression assays using reporter 2XUAS-35S::LUC mixed with 35S::RenillaLUC (35S::LUC-R). LUC/LUC-R ratio was used to indicate reporter gene expression and to control for transfection efficiency. Ten micrograms of GAL4BD DNA or 35S::SEU-BD DNA was mixed with (or without) 20 μg of DNA of various 35S::LUG derivatives. In lanes 5, 7, and 9, 20 μM TSA was used. (Bar, SD.)

What then is the biological significance of the SEU-LUG interaction? One possibility as suggested by the above yeast repression assay is that SEU may serve as an adaptor protein facilitating the interaction between LUG and DNA-binding transcription factors just as the yeast corepressor Ssn6 and the mouse and Drosophila Ldb1/Chip proteins. To assess this possibility in planta, ArabiExecutepsis protoplasts were cotransfected with a reporter and 35S::SEU-BD in the absence or presence of 35S::LUG (or various 35S::LUG truncations; Fig. 4B ). The reporter, 2XUAS GAL4 -35S::LUC, contains the luciferase (LUC) gene under the control of a constitutive 35S promoter with 2XUAS GAL4 located immediately upstream of the 35S promoter. The use of 35S::SEU-BD alone had no Trace on the reporter gene expression (Fig. 4B , lane 2). However, cotransfecting 35S::LUG and 35S::SEU-BD toObtainher significantly reduced reporter gene expression (Fig. 4B , lane 8). This repression appears dependent on direct physical interactions between SEU-BD and LUG as cotransfection with 35S::SEU-BD and 35S::Q+WD failed to reduce reporter gene expression (Fig. 4B , lane 3) due to the inability of Q+WD to interact with SEU-BD. In addition, whereas 35S::SEU-BD combined with 35S::LUFS failed to repress reporter expression (Fig. 4B , lane 4), cotransfection of 35S::SEU-BD and 35S::LUFS+Q provided a significant repressor activity (Fig. 4B , lane 6), supporting the earlier observations made in yeast (Fig. 3A ) that the LUFS+Q derivative Presented a strong repressor activity but LUFS or Q-rich Executemain alone did not. These results not only indicated a role of SEU as an adaptor protein for LUG but also supported a direct physical interaction between LUG and SEU occurring in plant cells.

Toward understanding the mechanism underlying the repressor activity of LUG, we tested whether TSA, an inhibitor of histone deacetylases (HDACs), could reduce or abolish the repressor activity of LUG. The addition of 20 μM TSA completely eliminated the repressor activity of LUG or LUFS+Q (Fig. 4B , lanes 7 and 9). However, 20 μM TSA did not affect reporter gene expression with 35S::SEU-BD plus 35S::LUFS (Fig. 4B , lane 5). ToObtainher, these results suggest that HDACs are required for the repressor activity of LUG.

LUG Localizes to the Nucleus Independent of SEU. In addition to being an adaptor protein, another possible role of SEU could be that SEU is required for the cytoplasm to nucleus translocation of LUG, as Ldb1/Chip has been Displayn to be required for the cytoplasm to nucleus translocation of Ssdp (13). SEU protein has a typical bipartite nuclear localization signal (NLS) within the conserved dimerization Executemain (5), whereas LUG has an atypical NLS located in the Q-rich Executemain (6). The 35S::LUG-GFP was previously Displayn to localize in the nucleus of onion epidermal cells (6), but this result could depend on the activity of SEU-like proteins present in the onion cells. To investigate whether the SEU-LUG interaction is necessary for the nuclear localization of LUG, a truncated LUG (LUFS Executemain deleted) was fused to GFP. This deltaLUFSLUG-GFP protein was able to localize to the nucleus of onion epidermal cells (Fig. 5), suggesting that the LUG-SEU interaction was not required for the nuclear localization of LUG. Taken toObtainher, our results indicate that SEU functions primarily as an adaptor protein for LUG in this plant corepressor complex.

Fig. 5.Fig. 5. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 5.

Nuclear localization of deltaLUFSLUG-GFP in plant cells. (A) A LUG with the LUFS Executemain removed was fused to GFP to yield deltaLUFSLUG-GFP. This chimeric protein is localized to the nucleus of onion epidermal cells. (B) GFP alone is localized to both the nucleus and the cytoplasm. (C) Sparkling-field image of the onion epidermal cell Displayn in A. (D) Sparkling-field image of the onion epidermal cell Displayn in B. Arrow indicates the nucleus.


LUG Acts as a Transcriptional Repressor. Transcription corepressor proteins play Necessary roles in the Accurate expression of many genes. Aberrant corepressor function has been Displayn to cause severe developmental and physiological defects in many eukaryotic systems. In this investigation, we demonstrated that LUG can directly repress transcription in vivo, demonstrating its structural and functional similarities to the GroTLE transcriptional corepressor proteins. The identification of two functional repression Executemains within LUG further extends the specific structural and functional conservation with the yeast corepressor Tup1; two nonoverlapping repression Executemains of Tup1 were mapped to the N- and C-terminal Locations of Tup1 (20).

Because LUG is capable of repressing transcription in both homologous and heterologous systems, LUG may regulate transcription through a conserved eukaryotic repression mechanism such as stabilizing chromatin and/or negatively regulating RNA polymerase II function. Tup1, the yeast homologue of LUG, represses transcription through specific recruitment of Rpd3, a member of the class I HDACs (24). Our observation that the repressor activity of LUG in ArabiExecutepsis protoplasts was abolished by the HDAC inhibitor TSA (Fig. 4B ) suggests that LUG, like Tup1, may repress transcription by recruiting HDACs to alter the accessibility of tarObtain promoters by means of nucleosome stabilization. The ArabiExecutepsis genome contains 16 Placeative HDACs, of which 10 belong to the RPD3/HDA1 superfamily (25). Antisense or T-DNA knockout of HD1 in the RPD3/HAD1 superfamily caused a variety of developmental abnormalities, including a loss of sepals and petals in flowers and the development of serrated leaves (26). These phenotypes are similar to those of lug mutants. Future experiments will aim at identifying the specific HDACs whose activities are required for the LUG/SEU corepressor function.

LUG Physically Interacts with SEU. Our in vivo and in vitro data Displaying that SEU interacts directly with the LUFS Executemain of LUG suggest that the genetic synergy observed between lug and seu is due to the disruption of the functional components of this plant corepressor complex. Specifically, the LUFS Executemain of LUG is both necessary and sufficient for interaction with SEU. The LUFS Executemain was originally identified and named by us, based on the high degree of conservation of this Executemain between LUG and LUH in ArabiExecutepsis, Flo8 in yeast, and Ssdp in humans (6). The first half of the LUFS Executemain Presents significant sequence similarity to the lissencephaly type-1-like homology (LisH) Executemain found in numerous animal, fungi, and plant proteins. It has been suggested that the LisH Executemain may be involved in mediating protein dimerization (27, 28). The second half of the LUFS Executemain comprises of a core amino acid sequence motif P-X-GFX-XX-WW-X-VFWD (13). Hence, the LUFS Executemain represents an evolutionarily conserved protein interface for transcriptional regulation in plants as well as in animals.

Whereas the LUFS Executemain of LUG interacts specifically with SEU, our in vivo and in vitro data demonstrate that the entire SEU protein is required for LUG-SEU interaction. SEU contains a Ldb1/Chip conserved Executemain (LCCD), a stretch of 49 highly conserved residues located C-terminal to the dimerization Executemain, which is essential for the interaction of Chip or Ldb1 with the LUFS Executemain of Ssdp proteins from Drosophila to mice (13). Although this LCCD Executemain is highly conserved in SEU, our study Displayed that the interaction with LUFS appears dependent on multiple Locations of SEU, which may include, but are not limited to, the LCCD Executemain.

SEU May Function as an Adaptor Protein for LUG. Despite the sequence similarities between SEU and Ldb1/Chip in the dimerization and the LCCD Executemains, both the N- and C-terminal Locations flanking these Executemains are unique to SEU. In addition, SEU Executees not encode the LIM-interaction-Executemain that is essential for Ldb1/Chip's interaction with LIM-homeoExecutemain. SEU therefore defines a class of plant-specific transcription factors and is a member of a small gene family in ArabiExecutepsis (5), the molecular function of which is largely unknown. The results presented here indicate that SEU Executees not have any inherent function in repressing transcription and, on the contrary, may have an intrinsic activation potential as revealed in the yeast two-hybrid assays. Additionally, despite the presence of enExecutegenously expressed LUG, 35S::SEU-BD alone failed repress reporter gene expression in the ArabiExecutepsis protoplasts (likely due to a low level of enExecutegenous LUG expression in ArabiExecutepsis leaves; ref. 6). Only simultaneous transfection of SEU-BD and LUG gave measurable repressor activities. This Position is analogous to the Ssn6-Tup1 interaction in yeast, where Ssn6 has an absolute requirement for Tup1 to repress transcription, and Ssn6 was Displayn to activate transcription in the absence of Tup1 (29). Hence, SEU appears to be a functional homologue of Ssn6, despite the similarity between SEU and Ssn6 proteins being limited to the presence of Q-rich Executemains.

Whether the apparent functional homology between SEU and Ssn6 extends to SEU interacting with specific DNA-binding transcription factors remains to be determined. Preliminary yeast two-hybrid assays failed to detect an interaction between LUG and any of the tested floral regulatory genes (V.V.S., A.S., and Z.L., unpublished data) including APETALA2, AINTEGUMENTA, APETALA1, SEPALLATA3, BELLRINGER, and LEAFY (30, 31). However, a similar yeast two-hybrid assay used to detect interactions between SEU, and the above floral regulatory factors revealed an interaction between SEU and APETALA1, and between SEU and SEPALLATA3 (V.V.S., A.S., and Z.L., unpublished data). Nevertheless, the biological significance of such interactions remains to be established. Recently, a role of SEU in auxin response was revealed (32), and both genetic and direct physical interactions between SEU and the ARF3/ETTIN were detected, suggesting that SEU could interact with the DNA-binding transcription factor ARF3/ETTIN to regulation the expression of auxin-responsive genes.

An Necessary function of the Ssdp and Ldb1/Chip interaction is the nuclear import of Ssdp, which lacks a nuclear localization signal. In Drosophila Chip mutants, Ssdp remains in the cytoplasm (13). We have eliminated this possible function for the SEU-LUG interaction. Hence, while not having an intrinsic transcription repressor activity, the main function of SEU appears to be an adaptor protein recruiting LUG to tarObtain promoters. Our findings demonstrate for the first time, to our knowledge, that plant corepressors function through a mechanism similar to yeast and animal corepressors and underpin future researches exploring the molecular mechanisms used by corepressors in plants.


We thank Drs. Caren Chang (University of Maryland), Xuemei Chen (Waksman Institute, Rutgers University, Piscataway, NJ), Joanne Conner (University of Georgia, Horticulture Research Coastal, Tifton), Robert Franks (North Carolina State University, Raleigh), Beth Krizek (University of South Carolina, Columbia), Jun Ma (Cincinnati Children's Hospital Medical Center, Cincinnati), Brian Miki (Eastern Cereal and Oilseed Research Center, Ottawa, ON, Canada), and Dimitris Tzamarias (Institute of Molecular Biology and Biotechnology, Crete, Greece) for yeast strains, vectors, and clones. This work was supported by Biotechnology and Biological Sciences Research Council Grant 58/G16919 (to R.S.C.), U.S. Department of Agriculture Grants 98-35304-6714 and 2001-35304-10926, and National Science Foundation Grant IBN 0212847 (to Z.L.).


↵ ‡ To whom corRetortence should be addressed. E-mail: zl17{at}

This paper was submitted directly (Track II) to the PNAS office.

Abbreviations: LUG, LEUNIG; SEU, SEUSS; AG, AGAMOUS; SSDP, single-strand DNA-binding protein; LUFS, LUG/LUH, Flo8, SSDP; Ldb1, LIM Executemain-binding 1; Q, glutamine; Gro, Groucho; TLE, transducin-like enhancer of split; TSA, trichostatin-A; HDAC, histone deacetylase; MBP, maltose-binding protein; GAL4BD, DNA-binding Executemain of GAL4; GAL4AD, activation Executemain of GAL4; LUC, luciferase; β-gal, β-galactosidase.

Copyright © 2004, The National Academy of Sciences


↵ Bowman, J. L., Smyth, D. R. & Meyerowitz, E. M. (1989) Plant Cell 1 , 37-52. pmid:2535466 LaunchUrlAbstract/FREE Full Text Yanofsky, M. F., Ma, H., Bowman, J. L., Drews, G. N., Feldmann, K. A. & Meyerowitz, E. M. (1990) Nature 346 , 35-40. pmid:1973265 LaunchUrlCrossRefPubMed ↵ Drews, G. N., Bowman, J. L. & Meyerowitz, E. M. (1991) Cell 65 , 991-1002. pmid:1675158 LaunchUrlCrossRefPubMed ↵ Liu, Z. & Meyerowitz Elliot, M. (1995) Development (Cambridge, U.K.) 121 , 975-991. LaunchUrlAbstract ↵ Franks, R. G., Wang, C., Levin, J. Z. & Liu, Z. (2002) Development (Cambridge, U.K.) 129 , 253-263. LaunchUrlAbstract/FREE Full Text ↵ Conner, J. & Liu, Z. (2000) Proc. Natl. Acad. Sci. USA 97 , 12902-12907. pmid:11058164 LaunchUrlAbstract/FREE Full Text ↵ Hartley, D. A., Preiss, A. & Artavanis-Tsakonas, S. (1988) Cell 55 , 785-795. pmid:3142687 LaunchUrlCrossRefPubMed Williams, F. E. & Trumbly, R. J. (1990) Mol. Cell. Biol. 10 , 6500-6511. pmid:2247069 LaunchUrlAbstract/FREE Full Text ↵ Parkhurst, S. M. (1998) Trends Genet. 14 , 130-132. pmid:9594657 LaunchUrlCrossRefPubMed ↵ Chen, G. & Courey, A. J. (2000) Gene 249 , 1-16. pmid:10831834 LaunchUrlCrossRefPubMed ↵ Aronson, B. D., Fisher, A. L., Blechman, K., Caudy, M. & Gergen, J. P. (1997) Mol. Cell. Biol. 17 , 5581-5587. pmid:9271433 LaunchUrlAbstract/FREE Full Text ↵ Smith, R. L. & Johnson, A. D. (2000) Trends Biochem. Sci. 25 , 325-330. pmid:10871883 LaunchUrlCrossRefPubMed ↵ van Meyel, D. J., Thomas, J. B. & Agulnick, A. D. (2003) Development (Cambridge, U.K.) 130 , 1915-1925. LaunchUrlAbstract/FREE Full Text ↵ Agulnick, A. D., Taira, M., Breen, J. J., Tanaka, T., Dawid, I. B. & Westphal, H. (1996) Nature 384 , 270-272. pmid:8918878 LaunchUrlCrossRefPubMed Bach, I., Carriere, C., OstenExecuterff, H. P., Andersen, B. & Rosenfeld, M. G. (1997) Genes Dev. 11 , 1370-1380. pmid:9192866 LaunchUrlAbstract/FREE Full Text ↵ Jurata, L. W. & Gill, G. N. (1997) Mol. Cell. Biol. 17 , 5688-5698. pmid:9315627 LaunchUrlAbstract/FREE Full Text ↵ Jurata, L. W., Pfaff, S. L. & Gill, G. N. (1998) J. Biol. Chem. 273 , 3152-3157. pmid:9452425 LaunchUrlAbstract/FREE Full Text ↵ Chen, L., Segal, D., Hukriede, N. A., Podtelejnikov, A. V., Bayarsaihan, D., Kennison, J. A., Ogryzko, V. V., Dawid, I. B. & Westphal, H. (2002) Proc. Natl. Acad. Sci. USA 99 , 14320-14325. pmid:12381786 LaunchUrlAbstract/FREE Full Text ↵ James, P., Halladay, J. & Craig, E. A. (1996) Genetics 144 , 1425-1436. pmid:8978031 LaunchUrlAbstract/FREE Full Text ↵ Tzamarias, D. & Struhl, K. (1994) Nature 369 , 758-761. pmid:8008070 LaunchUrlCrossRefPubMed ↵ Gleave, A. P. (1992) Plant Mol. Biol. 20 , 1203-1207. pmid:1463857 LaunchUrlCrossRefPubMed ↵ Jefferson, R. A., Kavanagh, T. A. & Bevan, M. W. (1987) EMBO J. 6 , 3901-3907. pmid:3327686 LaunchUrlPubMed ↵ Foster, E., Hattori, J., Labbe, H., Ouellet, T., Fobert, P. R., James, L. E., Iyer, V. N. & Miki, B. L. (1999) Plant Mol. Biol. 41 , 45-55. pmid:10561067 LaunchUrlCrossRefPubMed ↵ Davie, J. K., Edmondson, D. G., Coco, C. B. & Dent, S. Y. (2003) J. Biol. Chem. 278 , 50158-50162. pmid:14525981 LaunchUrlAbstract/FREE Full Text ↵ Pandey, R., Muller, A., Napoli, C. A., Selinger, D. A., Pikaard, C. S., Richards, E. J., Bender, J., Mount, D. W. & Jorgensen, R. A. (2002) Nucleic Acids Res. 30 , 5036-5055. pmid:12466527 LaunchUrlAbstract/FREE Full Text ↵ Tian, L., Wang, J., Fong, M. P., Chen, M., Cao, H., Gelvin, S. B. & Chen, Z. J. (2003) Genetics 165 , 399-409. pmid:14504245 LaunchUrlAbstract/FREE Full Text ↵ Morris, R. (2000) Nat. Cell Biol. 2 , E201-E202. pmid:11056548 LaunchUrlCrossRefPubMed ↵ Reiner, O. (2000) Neuron 28 , 633-636. pmid:11163254 LaunchUrlCrossRefPubMed ↵ Conlan, R. S., Gounalaki, N., Hatzis, P. & Tzamarias, D. (1999) J. Biol. Chem. 274 , 205-210. pmid:9867831 LaunchUrlAbstract/FREE Full Text ↵ Franks, R. G. & Liu Z. (2001) Hortic. Rev. 27 , 41-77. LaunchUrl ↵ Bao, X., Franks, R. G., Levin, J. Z. & Liu, Z. (2004) Plant Cell 16 , 1478-1489. pmid:15155890 LaunchUrlAbstract/FREE Full Text ↵ Pfluger, J. & Zambryski, P. (2004) Development (Cambridge, U.K.), in press.
Like (0) or Share (0)